Matching images and music for expressing emotions and music video generation

The Hong Kong University of Science and Technology 
Department of Computer Science and Engineering 

Final Year Thesis Oral Presentation

Title: "Matching images and music for expressing emotions and
        music video generation"

by 

Mr. Jason Shing Kit CHAN


Abstract: 

Expressing mood by sharing music is more understanding than sharing texts, 
however, it is hard to tell which music clip matches someone's mood. This 
can be solved if a song can be automatically suggested by only taking an 
image describing their moods. This paper describes an algorithm for 
matching images and music clips and generating music videos for an album 
by their emotional semantics. The algorithm uses support vector machine to 
cluster the training data by multiple type of emotion feature extracted 
and use the training result to build an emotion classifier used to match 
images and music clips.


Date                      : 26 April 2016 (Tuesday)

Time                      : 2:00pm to 3:00pm

Venue                     : Room 5506 (lift 25/26)

Advisor                   : Prof. Andrew HORNER

2nd Reader                : Dr. Xiaojuan MA